KR20170057202A - Semi transparent mark, a method for composing and detecting semi transparent mark,transparent mark and a method for composing and detecting transparent mark - Google Patents

Abstract

The present invention relates to a method for generating a semitransparent QR code by synthesizing a QR code image with an ordinary color background image, and a method for detecting a QR code by separating the QR code and an ordinary color image from an image obtained by capturing the semitransparent QR code. Accordingly, the present invention can use the QR code as a marketing means or a marker of augmented reality or virtual reality.

Description

Technical Field [0001] The present invention relates to a translucent mark, a semi-transparent mark synthesis and detection method, a transparent mark, and a transparent mark synthesis and detection method. BACKGROUND ART < RTI ID = 0.0 >

The present invention relates to a technique for translucently or transparently synthesizing a QR code image on a color background image.

QR code, smart phone, mobile device, camera, image processing, wearable computer, eyeglass display, virtual reality, augmented reality, mark

The present invention relates to the improvement of visual marks such as QR codes. A QR code is a kind of two-dimensional bar code, for example, a URL of a URL associated with an advertisement photograph of a magazine is printed by a QR code, and the user can take a picture with the camera of the smart phone and recognize the URL as a QR code. Such a QR code may be used for synthesizing an augmented reality image by detecting the attitude of the camera. Such existing QR codes are opaque, which hinders background images. In order to minimize the occlusion of the background image, the conventional QR code is printed small on a part of the background image, so that the user has to approach the QR code in order to photograph the QR code with the camera. In order to shoot a QR code from a distance, it is necessary to design a large QR code, but there is a problem that it hides the background image much.

 It is an object of the present invention to provide a semi-transparent or transparent QR code capable of designing a large QR code by composing a QR code in a translucent or transparent manner on a background image and also capable of viewing a background image and an image processing means .

In order to accomplish the above object, a translucent or transparent mark according to the present invention comprises a blue QR code synthesized on the basis of a brightness or a contrast of an ordinary color background image, And the brightness of the blue component image is adjusted to be larger than the brightness variation of the blue component image. A method of detecting such a semi-transparent or transparent QR code is characterized by detecting a QR code by binarizing a blue component image of an image obtained by capturing such a translucent or transparent QR code. Also, in order to stably perform such binarization, the image obtained by darkening the red and green component images in the blue component image is calibrated and binarized to detect the QR code before binarization.

By using the translucent or transparent mark according to the present invention, the QR code can be greatly attached to the surrounding environment so that it can be used as a marketing means or as a marker for augmented reality or virtual reality.

1 is an explanatory diagram for synthesizing two images
Figure 2 shows the change in the pixel value histogram
Figure 3 is a generalized version of Figure 2

Example 1

The object of the present invention is to provide a translucent qr code in which a monochrome QR code is semi-transparently synthesized on a color background image in order to solve the problem that a conventional qr code is opaque and obscures a background image. Specifically, a method of generating a translucent qr code image according to the present invention is as follows.

(Step 1-1) The ordinary color background image is decomposed into the three-color component image. That is, three gray images (red component image, green component image, and blue component image) are generated from the color image. For example, the gray image may be a 256-step image having the darkest pixel value of 0 and the brightest pixel value of 255. [

(Step 1-2) Black and white binary qr Change the white pixel value of the code image to 64 and the black pixel value to -64. The pixel range of the converted image is -64 to 64. The range (brightness difference) of such pixel values is referred to as contrast in the present invention. That is, the size of the pixel distribution range is 128. It is preferable that the absolute value of the value (-64, 64) is the same and the sign is different.

(Step 1-3) If the difference B (the distribution range size of the pixel value, that is, the contrast) between the darkest pixel value and the brightest pixel value of the blue component gray image obtained in Step 1-1 is smaller than the qr code Convert the pixel value to be smaller than the pixel range size Q (128) of the image but adjust it to B + Q <255. That is, the sum of the contrasts of the two images should be less than 255, which is the maximum possible contrast of the image. This way, the synthesized image can be output to the display or printed. The meaning of the above condition B + Q <255 means that the pixel range size of the composite image of the gray component image of the normal image and the qr code image should be smaller than 255. For example, if the darkest pixel value of the bluish gray image is 0 and the brightest pixel value is 255, the new pixel value multiplied by 60/255 is used as the pixel value of the darkened new image. Multiplying all the pixels by this constant (60/255) will darken the whole image and darken the original color image. In order to maintain the brightness of the original color image as much as possible during the conversion process of this step, the average pixel value of the blue component gray image is obtained, and a pixel having a larger pixel value is reduced in brightness and a pixel having a pixel value smaller It is desirable to increase the pixel value. Fig. 2 shows a histogram of the pixel values of the images before and after the conversion. In Fig. 2, the pixel range of the blue component gray image before conversion is 0 to 255 and the average pixel value is 164. When this image passes through steps 1-3, the image is converted into an image having a pixel range of 164-30 to 164 + 30. Fig. 3 is a generalized view of Fig. 3 shows a histogram of pixel values of an image in which an image having a pixel value range of f1 to t1 and an average pixel value of a is changed to an image having a pixel value range of f2 to t2. The arbitrary pixel value p1, q1 before conversion is converted into p2, q2, respectively, and the concrete value thereof can be obtained from the following proportional expression. (Non-proportional nonlinear transformations may be used, where the proportional equations are examples of possible transformations).

a - f1: p1 - f1 = a - f2: p2 - f2

t1 - a: q1 - a = t2 - a: q2 - a

In this way, by limiting the distribution range (contrast) of the pixel value of the blue component image of the background image to be smaller than the distribution range (contrast) of the pixel value of the qr code image, the image obtained by the camera is binarized to detect the qr code image If the distribution range (for example, 60) of the pixel value of the blue component image of the background image is smaller than the distribution range (for example, 128) of the pixel value of the qr code image, the step 1-3 is passed .

These concrete numbers 255, 128, and 60 are values for convenience of explanation, and values may be different when actually implementing the present invention.

Other algorithms other than the proportional equation described above may be used.

In this way, it is possible to reduce the contrast of the background image by using the same proportional expression or algorithm for the entire background image, but it can also be performed using a proportional expression or algorithm that is most suitable for each pixel of the background image. For example, in a combined image of a blue component image of a background image and a qr code image, the contrast of the border of the cell of the qr code is made larger than the contrast of the border of the background image within a certain region near the borderline The contrast of the boundary line can be reduced.

Here, the cell of the qr code means the smallest unit rectangular area of the same pixel value indicating the information of the qr code. It is preferable that a certain area in the vicinity is equal to or larger than the size of the cell of the qr code (the smaller of the lengths of the lateral or longitudinal sides of the square of the cell). When the contrast is adjusted as described above, binarization can be facilitated by examining the histogram of the pixels in the region of the size of the cell of the qr code around the pixel when binarizing each pixel of the synthesized image.

(Step 1-4) Obtain an addition image of the qr code image (image having a pixel range size of 128) obtained in Step 1-2 and the wave component image (image having a pixel range size of 60) obtained in Step 1-3.

Here, the addition image of the two input images means an image in which the value obtained by adding the pixel values of the pixels at the same position (x, y) corresponding to the two input images is the pixel value of the corresponding identical position (x, y) of the addition image.

(Step 1-5) The distribution of the pixel values of the addition image obtained in the above step 1-4 is adjusted so as to be within a printable range (for example, 0 to 255) or printable range on the display.

For example, if the pixel range of the wave component image obtained in step 1-3 is (255-60) ~ 255 and the pixel range of the qr code image is -64 ~ 64, then the range of the synthesized additive image (255-60-64) to (255 + 64). Since the maximum brightness of the gray image is 255 and the maximum value of the synthesized pixel value (255 + 64) is out of the range, a proper constant must be added to all the pixels of the synthesized addition image to shift the range of the pixel. That is, add -64 to all pixels of the sum image so that 255 + 64 is equal to or smaller than 255.

For example, if the pixel range of the wave component image obtained in step 1-3 is 0 ~ 60 and the pixel range of the qr code image is -64 ~ 64, then the range of the synthesized image is 0-64 ~ (60 + 64). Since the minimum brightness of the gray image is 0 and the minimum value (0-64) of the synthesized pixel value is out of the range, a suitable constant must be added to all the pixels of the addition image to translate the pixel range. That is, add 64 to all pixels so that 0-64 is greater than or equal to zero.

 (Step 1-6) A new color image composed of the red component image and the green component image generated in Step 1-1 and the new wave component image synthesized in Step 1-4 is generated. Then, the new color image is a translucent qr code image according to the present invention.

In this case, if the green component image or the red component image is excessively bright or the contrast is strong, it is advantageous to detect the qr code in the synthesized image after reducing the brightness or contrast. Further, when the contrast is excessively large in a specific part of the blue component image obtained in step 1-1 (i.e., when a too bright area and a too dark area are in contact with each other), it is also possible to further reduce the contrast only in that area. For example, only that part can be replaced with a histogram smoothed image. In this case, the histogram smoothing algorithm may use contrast-limited adaptive histogram equalization (CLAHE), https://en.wikipedia.org/wiki/Adaptive_histogram_equalization , or a similar method. Instead of histogram smoothing, the contrast of the image may be changed through other image processing such as gaussian blur.

This translucent qr code image can be printed on a paper in color or output to a color display such as an LCD or a beam projector, and the output image is photographed with a camera of a smartphone, and the photographed image is analyzed in a smartphone image processing program Qr code image and background image can be detected as follows.

(Step 2-1) A semi-transparent qr code color image is photographed with a color camera.

(Step 2-2) The captured color image is decomposed into three primary color component images. That is, a red component image, a green component image, and a blue component image of the photographed image are generated.

(Step 2-3) A composite image is generated in which the value obtained by combining the pixel values of the red component image and the green component image generated in Step 2-2 (weighted average) is used as the pixel value of the new composite image. For example, if a pixel value whose position is (x, y) in the red component image is R (x, y) and a pixel value whose position is (x, y) in the green component image is G The pixel value at the position (x, y) in the synthesized image is R (x, y) * w1 + G (x, y) * w2. Where w1 and w2 are weighted average composite coefficients, 0 <= w1 + w2 <= 1. (* Means multiplication) It is preferable to make the weight of green larger because the sensitivity of green is the highest among the three primary colors. That is, it is preferable that w1 < w2.

Concrete w1 and w2 values are obtained by detecting the qr code through these steps (steps 2-1 to 2-6) using various w1 and w2 values, and selecting the w1 and w2 values with the best detection For example, a value of w1 = 1/6 and w2 = 1/3 can be used.

(Step 2-4) A difference image obtained by subtracting the image synthesized in Step 2-3 from the wave component image generated in Step 2-2 is obtained. Here, the difference image is a pixel value obtained by subtracting the pixel value of the corresponding two input images. For example, if the two input images are denoted as I1 and I2 and the differential image of the two images is denoted as D, the pixel value D (x, y) of the differential image D at position (x, y) (X, y) -I2 (x, y) obtained by subtracting the pixel value of I2 from the position (x, y) The reason for this difference is that the blue color filter of the image sensor not only passes the blue light but also passes some red light and a little green light so that the red component of the captured image and the red This is to erase the object image (noise). Or because the camera's auto white balance function can cause the original color to change to a different color. Through such a difference, only the blue color image and the pure qr code image combined with the blue color image remain.

(Step 2-5) The difference image obtained in Step 2-4 is binarized to generate a qr code image. In the difference image obtained in the step 2-4, the brightness range 60 of the blue object is not larger than the brightness range 128 of the qr code image in the translucent qr code generation steps (1-1) to (1-6) And binarize these images so that a qr code image can be obtained easily. Since these brightness range values (60, 128) are only one example for convenience of explanation, other values may be used in actual implementation.

(Step 2-6) In the difference image obtained in Step 2-4, the qr code image obtained in Step 2-5 is multiplied by an appropriate weight to obtain a subtracted difference image, which is obtained by subtracting the qr code from the blue Component image. When the color image of the blue component image is appropriately increased and the color image of the red and green component image obtained in the step 2-2 is synthesized, the background image is obtained by removing the qr code.

If the qr code image is not detected as a result of the steps (2-1) to (2-6), the change in brightness of the normal background image in step 1-1 is significant. In this case, it is necessary to adjust the brightness of the background image so that the brightness of the background image is small, and then combine the semi-transparent qr code image. For example, it is preferable that the translucent qr code image is synthesized by reducing the brightness variation width of the red and green component images of the normal color background image (step 1-1) through the same process as step (1-3).

In the above description, the qr code image is synthesized with the blue component image of the background image because the blue is the least sensitive to the human eye, so that the qr code is not conspicuous in the synthesized image. That is, when the human eye sees the synthesized image of synthesizing the qr code with the blue component image, the background image is similar to the original image, and the qr code image is the faintest.

An arbitrary other mark image may be used instead of the qr code image to be synthesized with the normal background image. For example, an image such as a logo, an icon, a marker for augmented reality, a barcode, or a letter may be synthesized by using it instead of a qr code image.

The synthesized image is analyzed by the above method to detect a marker or a qr code, and a three-dimensional relative position and direction between the camera and the qr code or marker is calculated, and the video object video see through) augmented reality system. At this time, it is preferable to delete the qr code or marker from the photographed image and synthesize the virtual object. The user can see a more natural virtual reality image since the qr code or markers are not visible in the background of the synthesized virtual object image. Such a system may be configured using a camera of a smart phone or may be configured using a camera attached to a spectacular display.

To use these blue qr codes or markers in an optical see through augmented reality system (such as Microsoft's holographic lens system or Google Glass), you can use a spectacular display You can attach a yellow filter (a filter that passes red and green and blocks blue). Do not attach a yellow filter in front of the camera. That is, the yellow filter should be installed only in front of the wearer of the eyeglass type display. The yellow filter serves to block the blue component of the qr code image. At this time, it is preferable to use a filter that passes a little blue color rather than full yellow color, so that a more natural color can be seen.

It is preferable that such a yellow filter eyeglass is a variable filter that is normally transparent and changes color to yellow when a blue qr code is detected in an image taken by a camera. Such a variable color filter can be implemented, for example, as an LCD-type transparent display. It is also preferable that the portion changing to yellow color changes only the area occupied by the qr code in the wearer's field of view.

Such a blue qr code can be attached to a wall, a ceiling and a floor of the room, for example, and augmented reality can be implemented in the entire room by wearing a spectacular display. In addition, a robot camera such as a cleaning robot recognizes the qr code and recognizes its own position so that it can run autonomously. This blue qr code is less noticeable than the conventional black qr code, so it can alleviate visual discomfort to the naked eye. You can also hide the qr code by composing translucent blue qr code into a normal picture or picture.

Example 2

The specific color component image (for example, blue component image) of the translucent qr code image generated in the first embodiment is a composite image of the qr code image and the blue component image of the normal image. And other color component images (for example, red and green component images) other than the specific color component of such translucent qr code image are normal images. (Inverted image, negative image) of the qr code image used in the translucent qr code image (which is referred to as a second image in this embodiment) is converted into a specific color component image (for example, a blue component image) (This is referred to as a first image in this embodiment). The pixel values of the remaining color component images (for example, red and green component images) of the specific color component image of the first image are all zero. That is, the first image is a blue image, and the content thereof is a qr code that is inverted in black and white. If the first image and the second image are alternately displayed on the display, a person can not see the qr code image and can only view the normal image. However, when such a display is photographed with a high-speed camera, the qr code image can be photographed and the qr code can be detected by the image processing means.

The method of detecting the qr code in the photographed image is the same as that described in the first embodiment. In other words, the qr code image can be obtained by binarizing the wave component image of the photographed image.

The qr code, which is not visible to such a person, can be output to a TV screen so that the user can photograph the qr code on a smart phone to obtain information related to the program currently being broadcast. For example, you can get information related to the products that are displayed on the current screen. Each frame image of the original moving image to be outputted on the TV screen is combined with the qr code every moment as a background image, and output to the TV screen, so that the augmented reality image can be synthesized on the basis of the moving image. This is superior to the existing augmented reality image with still image as background. For example, an image in which a shark in a TV jumps out of the screen may be synthesized into an augmented reality image.

This embodiment is similar to the 'Stereo image-based image processing system' of Korean Patent Registration No. 10-2012-0109581, which is a prior patent of the inventor of the present invention, but the mark of this embodiment has an advantage that it is invisible to the human eye. On the other hand, the technique of Registration No. 10-2012-0109581 has a problem in that the mark image normally covers the normal image because the mark image and the normal image are outputted alternately.

Example 3

The first image and the second image of the second embodiment can be output to the right eye and the left eye, respectively, on the three-dimensional display of the polarization scheme.

The polarizing type three-dimensional display has a display screen divided into a plurality of slender horizon lines, and different polarizing filters are attached to the even-numbered areas and the odd-numbered areas, respectively. In this display, the image to be displayed in the right eye is output to the even-numbered region, the image to be displayed in the left eye is outputted in the hole-numbered region, and different polarizing filter glasses are attached to the left and right eyes, can see.

Another example of a polarized 3d display is a beam projector for use in 3d theaters. These beam projectors project images to be displayed on the left and right eyes of the stereo image alternately with different polarizations on the screen. The first image and the second image of the second embodiment can be alternately projected on the screen with different polarizations through the beam projector.

It is possible to easily detect the qr code of the first image by attaching a polarizing filter which passes only the polarized light of the first image in front of the camera lens and photographing the 3d screen of this polarization method. A person who does not wear polarized glasses can see both the first image and the second image and can not see the qr code image. That is, a person who does not wear polarized glasses can only see the original normal image without the qr code.

OI: Normal color image
QI: QR code image
4Q: Synthetic translucent QR code image

Claims (7)

In a translucent mark image generation apparatus including a mark image and a background image
If the contrast of the border line of the mark image is
Contrast adjusting means for adjusting the contrast of the background image to be larger than the contrast of the boundary line of the specific color component image;
When the contrast is adjusted
An addition image generating means for generating an addition image of a specific color component image of a mark image and a background image;
The color component image of the addition image and the background image is converted into a color component image
And a color image generating means for generating a color image, wherein a boundary line of a specific color component image of the background image is a boundary line distributed within a critical distance from a boundary line of the mark image
Wherein the critical distance is at least half of a horizontal size or a vertical size of a cell at a border of a mark image, whichever is smaller. The apparatus of claim 1, wherein the specific color is one of three primary colors. The method of claim 1, wherein
The contrast of the mark image included in the remaining color component images other than the specific color of the translucent mark image is
The contrast of the normal image included in the remaining color component images other than the specific color of the translucent mark image
The remaining color component images other than the specific color of the semitransparent mark image do not include the mark image
Wherein the translucent mark image generating device comprises: Adjusting the contrast of the specific color component image of the background image to be not larger than the contrast of the mark image;
An additive image generation step of generating an additive image of a specific color component image and a mark image of the contrast-adjusted background image;
Adjusting the brightness or contrast of the addition image to a range that can be printed or displayed on a display;
A color image in which the brightness or contrast of the step is adjusted to a printable or displayable output range and the remaining color component images of the background image are color component images
Generating a color image;
And a translucent mark image. The method of claim 4, wherein
Wherein a maximum value and a minimum value of pixel values of the mark image are equal in absolute value and different in sign,
Translucent mark image synthesis method. Extracting a specific color component image of a translucent mark image including a mark image and a background image but including a mark image as a specific color component image
And binarizing the extracted specific color component image. The method of claim 7, wherein
The step of extracting the specific color component image
And obtaining an image obtained by subtracting an image obtained by darkening the color component image excluding the specific color component image from the specific color component image
Wherein the translucent mark detection method comprises:

Images